基于神经网络的PMLSM建模  

Modeling of PMLSM based on neural network

作　　者：张向文[1] 杨海柱[1] 王福忠[1] 焦留成[1] 

机构地区：[1]焦作工学院电气工程系,河南焦作454000

出　　处：《焦作工学院学报》2002年第2期136-139,共4页Journal of Jiaozuo Institute of Technology(Natural Science)

摘　　要：永磁直线同步电动机 (PMLSM)是直线电机驱动的垂直运输系统的核心 ,它的模型的建立对于整个系统的控制与运行分析具有重要的意义 .利用BP算法 ,用C语言编制程序 ,实现了一个单输入层单输出层 4隐含层的神经网络 ,并用实验数据对网络学习训练 ,建立了PMLSM的模型 .通过实验验证 ,该模型的输出结果与实际结果十分接近 ,能够反映PMLSM的基本运动特性 ,因此这种模型可以作为PMLSM的仿真模型进行控制和运行特性的仿真研究 ,并且在进行控制时可以把这种模型作为系统的辩识模型进行预测控制 ,使控制效果更好 .Permanent magnet linear synchronous motor(PMLSM) is the core of vertical transport system driven by linear motor.Its modeling has important significance to the control and performance analysis of the system.A neural network,including a single?input?layer,a single?output?layer and four hidden?layers,is realized by programing with C language based on BP algorithm.And the model is formed from training the network with experimental data.Its experimental verification shows that its output is very close to the experimental result,which can response the basical charactenistic of PMLSM.This model can be used as a simulation model to study PMLSM's performance,or to PMLSM's predication control as an identification model.In a word,the modeling of PMLSM will open a new direction for the study of PMLSM.

关 键 词：BP网络 PMLSM 神经网络 永磁直线同步电动机 电磁场 

分 类 号：TM351[电气工程—电机]